Elastic's business model?

I’m intrigued about Elastic. It’s pretty clear that the type of search that they specialize in has a long runway of growth. What I don’t understand is the open source business model. When I think about SaaS, the two biggest products that come to mind as exemplars would be Office 365 and Creative Cloud. When companies buy a subscription to Office 365, they’re getting access to the Software, and then Microsoft doesn’t need to do any more lifting to make that customer happy. Nobody’s calling Microsoft for support with Word and Excel. As close as I figure, Alteryx is the closest pure approximation here in terms of the smaller enterprise cloud companies we like here.

But Elastic is open source. Anyone can just grab it. In fact, Amazon has already created its own product based on the Elastic stack. So what is Elastic selling? I can’t seem to figure it out. What I worry is that they’re product offerings involve a lot of implementation and operation on their part on behalf of the customer. If this is the case, isn’t their service model more “service as a service” and less “software as a service”? I’d really like to open a position but I can’t get around this mental hurdle. It’s hard to glean too much from their finances since they just IPO’ed, but their operating cash flow isn’t improving comparing 2017 to 2018, which doesn’t indicate that the business is set up as I’ve described, but also doesn’t help me to feel that it isn’t.

Does anyone have any insight on this?


Hi bodhibob,

I’m no techie guru, but if you read through the company S-1 filing you can learn quite a bit, and I think find all the answers to your questions. They do offer an open source model, similar to MongoDB.

here is a snippet from the filing


Our origins are rooted in open source, which facilitates rapid adoption of our software and enables efficient distribution of our technology. Developers can download our software directly from our website for use in development and production environments. Since January 1, 2013, our products have been downloaded over 350 million times. These downloads include both free and paid products. Open source also fosters our vibrant community of developers who help improve our products and build on top of them. As of July 31, 2018, our community included over 100,000 Meetup members across 194 Meetup groups in 46 countries. Meetup members are individuals who opt into an Elasticsearch Meetup group on meetup.com, an independent third-party website.

Our business model is based on a combination of open source and proprietary software. Many features of our software can be used free of charge. Some are only available through paid subscriptions, which include access to specific proprietary features and also include support. Unlike some open source companies, we do not build a separate enterprise version of an original open source project. Instead, we develop and test one robust codebase, over which we maintain control. We believe that maintaining full control over the source code enables us to develop better products for our users and customers. Our sales and marketing efforts start with developers who have already adopted our software and then evolve to departmental decision-makers and senior executives who have broad purchasing power in their organizations. All of these actions help us build a powerful commercial business model on top of open source.

Our customers often significantly expand their usage of our products over time. Expansion includes increasing the number of developers using our products, increasing the utilization of our products for a particular use case, and applying our products to new use cases. We focus some of our direct sales efforts on encouraging these types of expansion within our customer base. We believe that a useful indicator of our customers’ tendency to expand their usage of our products is our Net Expansion Rate, which measures expansion in existing customers’ annual subscriptions over a twelve month period. Our Net Expansion Rate was 142% as of July 31, 2018 and over 130% at the end of each of our last seven fiscal quarters.

Its quite evident, at least to me - giving a “free taste” if you will, works very very well for Elastic as well as MongoDB. You look at the key metrics, revenue growth, retention rate, and growth in customers is proof enough at this stage, that the model is working, and the high retention rates show you that customers generally love the product so much that they buy more of it the next year. As the case is for many small companies growing like wildfire, they will plow every penny they can right back into Sales, Marketing, R&D, and forego making a profit or in this case positive cash flow. As they achieve greater scale we should start to see the company squeeze out positive cash flow before positive net income. But of course, we will follow their progress closely.



Some added info from GEEKWIRE, back two months ago when ESTC IPO’ed.


Another strong IPO for SaaS investors, as open-source search provider Elastic enjoys huge pop on opening day

“It started off as a free service, and then Elastic built additional features around the project that developers can pay to use as their use of the search engine grows. The company has also added products that use the basic search capabilities to help infrastructure administrators monitor logs across their servers, and that help security professionals search for malware or intrusions in their networks using log data.”

Other comments on IPO-day experience:
"Elastic’s stock opened at $70, nearly double the original asking price of $36 a share, and held that ground throughout the trading day, closing at $70.50. That values the company around $4.9 billion, and comes on an otherwise down day for the stock market.

“It is a bit of a surprise,” acknowledged Shay Banon, chairman and CEO of Elastic, in an interview with GeekWire. “We did come back from the road show with a good feeling with investors, our story resonated with them,” he said, but Elastic decided to price at $36 to preserve what Banon called a “sane” valuation multiple."

1 Like

Its quite evident, at least to me - giving a “free taste” if you will, works very very well for Elastic as well as MongoDB.

If it weren’t for Atlas I would not invest in MDB. My reasoning is based on the history of MySQL, the most widely used open source SQL database which I’ve been using for almost two decades without ever spending a penny on it. It comes free bundled with MacOS. In the end the owners of MySQL didn’t IPO but sold the company which eventually wound up as part of Oracle. Since lots of people hate Oracle they transitioned seamlessly to MariaDB, developed as open source by the same people who developed MySQL. I didn’t even notice until my database error messages came from MariaDB instead of from MySQL. My web hot hadn’t even bothered to announce the change, it was so seamless and problem free.

Now, Atlas changes all that and it’s already 22% of Mongo revenue.

Denny Schlesinger


Now, Atlas changes all that and it’s already 22% of Mongo revenue.

Would you mind elaborating on that for us non-techies, Denny?



The simplest way I can explain it is by example. I used to develop websites. I do the development on my home machine but the finished websites run on a server somewhere, mine currently in California. What this means is that my development (home) machine has a version of the same software that runs on my web host’s server. They don’t have to be identical, just compatible. currently I run MySQL at home and MariaDB on the web host’s server. I run MacOS, the host something else. I have a tiny operation and I pay $35 a month for the web hosting which includes the main server in California, an offsite backup somewhere in the midwest (we are not told where for security reasons) and an emergency server in Virgina. I contact the help desk maybe a dozen times a year, if that. Suppose I have to update phpMyAdmin on my development machine – the software that manages the database – and it takes me an hour. At programmer rates, how much would that cost? $100? $200? Three to six months of web hosting cost! Programmer’s time is very expensive. This is what makes Anything as a Service so attractive. Updating phpMyAdmin at the host probably takes less time because that’s what they do for a living and the cost is spread over hundreds of customers.

MongoDB’s original business based on open source software relies on what Duma calls “Freemium,” the open source is free but the premium part is paid. I bet that only a tiny part of all the MongoDB download users pay for the premium stuff which is why MySQL never became truly profitable. Having Mongo run the database for you on the cloud of your choice is a much better deal, it eliminates tons of complexity, like Pivotal says, “just do it, I don’t care how.” As long as Atlas works the client retention should be phenomenal. My CitiCard is 28 years old and my web host at least 15. They work, why change?

I hope that explains it. Mongo breaks out the revenue growth rate of Atlas which is much higher than the traditional Freemium revenue. At Elastic, check out the various growth rates. For a while Hortonworks was a Saul darling but it was not “Something as a Service.” It’s very difficult to make money on something that you give away for free.

Denny Schlesinger


I view Atlas similarly to Shopify Plus. The bulk of users of the lower tiered offerings are not meaningful to either SHOP’s or MDB’s move to profitability. Atlas and Shopify Plus are and where investors should focus upon.



This is the part of the S1 that you quoted that gets to the heart of my question:

“Some are only available through paid subscriptions, which include access to specific proprietary features and also include support.”

If the key benefit to having a paid subscription is the support that is offered, then Elastic is a service business, and not a SaaS business. Now, if the key benefit is more related to the proprietary features, and support is a marginal offering, this isn’t an issue and Elastic can indeed be considered a SaaS offering. But if the support is the key driver, it’s something different.

Under risk factors Elastic mentions their employee headcount a couple of times. Of course, a growing company needs to increase its headcount. But a company that relies on support personnel is going to need to increase its headcount much faster than a company that only needs to invest in developers.

From the S1:

“Our employee headcount and number of customers have increased significantly, and we expect to continue to grow our headcount significantly over the next year. For example, our total number of customers has grown from over 2,800 as of April 30, 2017 to over 5,000 as of April 30, 2018 to over 5,500 as of July 31, 2018. The growth and expansion of our business and offerings places a continuous significant strain on our management, operational, and financial resources.”

interesting that they break out their customer numbers but not their headcount numbers…


“As we expand our business and operate as a public company, we may find it difficult to maintain our corporate culture while managing our employee growth. Any failure to manage our anticipated growth and related organizational changes in a manner that preserves our culture could negatively impact future growth and achievement of our business objectives. Additionally, our productivity and the quality of our offerings may be adversely affected if we do not integrate and train our new employees quickly and effectively. Failure to manage any future growth effectively could result in increased costs, negatively affect our customers’ satisfaction with our offerings, and harm our results of operations.”

This appears to be standard boilerplate, but in light of my question I find it interesting that they are so explicit about the challenge of managing employee growth.

Also, Denny’s comments are interesting to me as this line of questioning appeared in my head after reading a negative blog post about MongoDB as an investment. So it’s curious that MongoDB comes up in this conversation as a related case. After reading the blog post linked below, I concluded that the author may have a point, but the point had been rendered moot through the introduction of Atlas, which follows Denny’s reasoning exactly. I asked myself if the author’s point, while no longer being relevant to MongoDB, might be applicable to Elastic in its current form. I couldn’t figure it out, so I brought it up here.

Blog post in question: https://medium.com/@honnibal/what-mongodb-tell-the-market-is…


We had 1,129 employees as of October 31, 2018. (from the 10Q)

With 5000 customers at the end of April, and 5500 at the end of July, and presumably 6000 at the end of October, they seem to me to be running a very employee-light business. I think you can pack your worries away. I’m afraid you were overthinking this. Probably 300 or 350 of those employees were in R&D, others in administration, more in sales and marketing, accounting, human resource management, etc etc etc, it doesn’t leave many to intensively service 6000 customers.




Perhaps there is some confusion of the term “support” when dealing with software licenses and subscriptions. Nearly all software has some level of support built into a license or a subscription. A software company would have to provide support for their product. They get gliches or a new feature rolls out that a customer is confused on or they can’t get something to function quite right.

This most certainly is included in a Microsoft 365 subscription.


Premium support via chat or phone with Microsoft experts.

Microsoft states in their 10-Q “We generate revenue by licensing and supporting an array of software products”.

This is a typical statement of all software companies.

Also typical of most software companies is a services department. This usually consists of consulting and training and certification. Elastic does have this “services” business as well. For some companies services is a big part of the business at least presently, like with Pivotal Labs.

Here is how Elastic describes their revenue generation.

We generate revenue primarily from sales of subscriptions for our software. We offer various subscription tiers that provide different levels of access to paid proprietary features and support. We do not sell support separately. Our subscription agreements for self-managed deployments typically have terms of one to three years and we bill for them annually in advance. SaaS customers may purchase subscriptions either on a month- to-month basis or on a committed contract of at least one year in duration. Subscriptions accounted for 92% and 93% of our total revenue in the six months ended October 31, 2018 and 2017, respectively. We also generate revenue from consulting and training services.

You’re paying for the propriety software features and getting support. Support is not sold separately. And the last sentence is talking about services.

Here is how Elastic breaks out these revenues and the percentages of the total business.

Self-Managed Subscriptions- 76%
License - 16%
Subscription - 60%
SaaS - 16%
Total Subscription- 92%
Professional Services- 8%

Consulting and training is a very small part of what they do. The “freemium” strategy is to get customers started and experimenting with Elastic and then to transition them as they need more features to fully realize the potential. But no, they are not primarily a services company by any definition. And with consistent top line growth hanging at greater than 70%, they are clearly succeeding in winning paying customers. Of course so long as that continues…



The benefits to having a paid subscription is primarily to get access to “proprietary/differentiated” features that is not part of the Elastic Open Source (OSS) product (or for example Amazon’s ES offering), as well as “support” in a production-type/live environment from the people who built the software and who know and understand it best.

To give you a tangible example, let’s say you somehow found yourself downloading/experimenting with the Elasticsearch OSS for a proof of concept (POC). You put some stuff (data) into it and you ask it some interesting questions (search your data) and you like what you see (your new application feature/functionality). Now it is time to grow your POC a bit and put more data into it (say billions or more of data that you care about). You then play with ES some more and eventually determine that it fits your needs and you are ready to take the next step and move to production/live. But turns out that you work for a company where security is a priority and is non-negotiable. You then find out that to secure ES, you will need the XPACK for security (a proprietary ES feature) which is only available from ES as part of the Gold or Platinum level paid subscription. So you love the solution but in order to get your company to approve ES and go to production, you will need to have the security feature which means you will need to subscribe (Gold or Platinum).

Above is just 1 teeny tiny example of what an Elastic subscription means. The Elastic family of products have many interesting and targeted use-cases that are unique where only a handful of solutions are available (today anyway). One thing to note about open source as it applies to Elastic is that I/you (the customer) play an active role in the open-source community and in deciding if an Elastic product makes sense to you or not. What this means is since majority of the Elastic products are open source (free to install and play with), I/you can do our own POCs on our own dime without any sales pitch from Elastic and I/you can arrive at our own conclusions about whether an Elastic product is crap or not. Not only that, if I/you find something that is lacking in an Elastic product, I/you can actually contribute to the product and make the product better. This is the essence of open-source as it applies to Elastic products. There is a community of “users” and they all contribute in one way or another, which then advances the product/s, which then benefits everybody in the community.

Furthermore, what this means is Elastic has no incentive to put out crappy products whatsoever since they can only convert me/you to a subscribing customer if I/you find that the product is actually “more than excellent” (and really, being good enough is not good enough here - the product has to be excellent to be able to convert me/you to a paying subscriber). As a result of this model, the customers that eventually convert are those that have done their due diligence (on their own dime) and already know that they “love” the product. By the time they knock on Elastic’s door, they are ready to sign and their discussions with Elastic is probably just to check off some boxes and some higher-up decisions.

If you have any more questions, please fire away and I will be glad to answer.


This is exactly the kind of information I was looking for. Thanks so much for helping me to understand. My new understanding is that although elastic is a “search” company, the search aspect is provided for free and the revenue generating product is the peripheral enterprise cloud products that allow a company to roll it out and keep it running. Now I get it. It’s not so different from Alphabet. Google search is provided for free and the revenue generating product, Google Ads is layered on top. This gets me excited about the company. Suddenly, in my mind I don’t see the company as experimenting with an unproven business model, but rather building on the existing search revenue paradigm.


Hi bly2k,

That was such a clear and useful and informed explanation of how Elastic works, and showed so much understanding and familiarity with the products, that I was just wondering if you are a developer working with the products yourself, or part of senior management at Elastic? :grinning:

For myself, I try not to get lost in the mechanics, but just follow the numbers, and if the company can grow their sales at 70% to 80% per year, that is enough to convince me that they know very well what they are doing.




Although “search” is indeed one of Elastic’s offering, they have evolved over the years to more than search. If you go to their website and browse previous “customer/user” presentations at their annual ElasticON conferences (this is where the community comes together to talk about the products as well as learn about the Elastic roadmaps), you will start to get an understanding of their production use-cases. But really search is just one part of Elastic (they started from search but they do many other things now).

IMO, one of the core/underlying differentiators that the Elastic products bring to the table is their ability to provide realtime insights on humongous volumes of data, and allowing I/you to do this in my/your own “basement” to see and decide for ourselves if the products are really enterprise scale - i.e. will it actually work with terrabytes/petabytes of data and will it give you your answers/insights in realtime (usually milliseconds). On top of this, I/you can actually contribute (Elastic is very open and they welcome contributions from the community in a significant way) and advance the products to help solve not only your own problems but potentially other people’s problems in the community ← this is the primary strength of the company - the community advances their products based on real world needs and the company kinda acts as the “leader” to get the community contributed features (as well as the proprietary features - which the company builds from from listening to the community) to an enterprise level which Elastic can/will then support. And since Elastic is so small (relatively speaking), critical enterprise level features sometimes move extremely fast and will allow the community to be able to take advantage of any advancements to the products and go to production much quicker.



I have worked with Elastic products for a while now (I believe since version 0.18 or thereabouts back in 2011). I have a good understanding of what can be done with their products (at an enterprise scale), what their strengths/weaknesses are, and really I am just a very happy user of Elastic products. I’ve met many other Elastic users and in general, their experiences mirror mine as well.

I will leave the number crunching to you guys since you are all better than me at that, but if you have any questions about Elastic or their products, I am more than happy to answer.



I understand Elastic’s competitors to be Splunk and Solr (you likely know more competitors). Can you comment on why a customer might choose one vs the other(s) ---- and if you feel there is one big winner or multiple winners in this space?

Also, if anyone can tell me how to search this board for older messages, I would appreciate. For example, I know there has been other research/posts on ESTC yet I can’t seem to search for “ESTC” or “Elastic” to return results from Saul’s board. Rather than clutter up the board, please e-mail me if you know how to do this.

Also, if anyone can tell me how to search this board for older messages, I would appreciate. For example, I know there has been other research/posts on ESTC yet I can’t seem to search for “ESTC” or “Elastic” to return results from Saul’s board.

Since this will be useful to anyone on the board I’ll post it here instead of off-board:

Enter the following on Google and do a search for whatever topic you want. (Here it’s Elastic)

site:fool.com “Saul’s Investing” Elastic


Just wondering whether Elastic might have an elegant solution for this search problem on Fool boards. ; )


Just wondering whether Elastic might have an elegant solution for this search problem on Fool boards. ; )

Yes they do. They acquired a company named swiftype a while back that does indexing and site search specifically for website content:


Or with Elasticsearch OSS, it’s actually not very difficult to do this. This is one use-case that ES was built for and is very good at.


I understand Elastic’s competitors to be Splunk and Solr (you likely know more competitors). Can you comment on why a customer might choose one vs the other(s) ---- and if you feel there is one big winner or multiple winners in this space?

This is a difficult question to answer. Apache Solr is closer to ES since Solr is also open source. Splunk is completely different (in terms of the underlying model anyway). Not sure if this helps but:

  1. ES is relatively still a baby here. You can actually google some info and find that ES adoption has been so fast in a very short period of time (for example comparing against Splunk). Splunk and Solr have been a around for a while now and are likely well entrenched in the industry.

  2. Splunk is a closed end system, i.e. you are sold a product where you do not see the “guts” of how it works, or get to customize it if you like/need. Some people like that, some people don’t. It’s a matter of taste, and also cost. Splunk charges by volume of data (last time I checked) so that could be the primary consideration for choosing/not choosing Splunk. I have seen many smaller startups shy away from that model simply because it can be costly for them. On the other hand, Splunk is fairly mature and contains a lot of features that ES/Solr are still likely catching up to. You can think of Splunk as the “de-facto” industry choice (for machine data anyway) but then if you come from an open-source background/roots, then you will likely be more open to considering ES/Solr.

  3. ES and Solr are both open-source and are closer to each other - in fact, the underlying low-level technology used by both ES and Solr are the same. Between these two, it’s probably a toss-up. My guess is companies who have user Solr for years (before ES came to the scene) will continue to use Solr. On the other hand, companies who are converting from another commercial enterprise search product, or companies newer to search will probably consider ES as it has a “modern” feel to practically everything in how it was designed and how you work with it. This is not to say that ES is a more mature product - it is actually less mature than Solr but since it is newer, ES was designed from the ground up to have “less friction” in terms of overall usage and cloud deployment. ← what this is translates to is that ES will likely have an “easier” learning curve since it was built with that in mind, and this will likely allow you to play with it and go to production faster (all things being equal).

  4. Solr is not “owned/led” by a single company. There is a primary company (among many) that supports it but they do not “lead” the product advancements. ES on the other hand is slightly different - it is open source but only the company (Elastic) will have the final say on whether something goes into the OSS product/s or not. So this means somebody actually leads the products and ensures that the quality of the contributions is up to a level that Elastic can/will support. Neither is good or bad - it’s just a matter of taste and alignment with your principles/preferences. Some people are more comfortable getting support with well-defined SLA’s from an actual “owning” company (like Elastic), but others don’t care much about that.

  5. Feature wise, you probably will not go wrong with any of the above. What it will likely boil down to is cost and whether you have roots in the open-source community or not.